The mobile computing revolution is definitively in full-swing: smartphone sales surpassed desktop sales in Q4 2010, and in the developing world, your computer is more likely something that you hold in your hand than on your lap.
The increase in mobile platforms isn’t limited to phones however: the Pebble smartwatch seems to be a trend-setter in a category which may soon be entered by Apple, and Google’s Glass, whether it succeeds in its current form or not, is a sign of a future of wearable computing. “Smart” devices are popping up everywhere, from the refrigerator, the thermostat, the TV, to even the deadbolt.
It doesn’t seem to be a stretch that the future of the computer lies everywhere around us, not just in a few specific devices.
But, as has always been the case with computing platforms, the success of a platform relies largely on the applications developed for it, and in many cases depends on a single “killer app.” The platform enables new functionality, it rarely provides that functionality itself. Apple’s advertising recognizes the importance of having “an app for that,” and Blackberry’s developer bounty puts a price on the importance to the ecosystem of having applications.
The speed and cost at which new applications can be deployed on these new computing devices dictates the rate of innovation for the hardware revolution, and the return to native apps is slowing that rate considerably.
With performant web apps, we were finally given a way to develop and deploy applications that were truly cross-platform . In addition to being cross-platform, web apps are managed almost entirely by the vendor - shortening cycle times for development, and lowering costs for the customer. The SaaS revolution speaks to the strength of the web app model for both software vendors and consumers.
The huge gold rush that came with the exploding popularity of smartphones led to a dramatic shift in developer attention away from web apps, and towards native mobile apps. Native apps are now top of mind for anyone thinking of starting a tech startup, and with every new platform that is released, there is a new set of developers champing at the bit for a chance to develop for it (see: Glassware).
As a result of this return to native applications, if you have a great idea for an application, you once again have to develop it for every platform individually. This type of platform fragmentation is an absolute nightmare for an undercapitalized technology company. Developing for the web is cheaper than ever, but the proliferation of new devices with new platforms promises to eliminate the gains the technology community has made in the accessibility of starting a company, and thus, the number of successful companies that are started.
The devices where this is most striking is in “Smart TV’s”. In the past 3 weeks I’ve hooked up two different TV’s to our local network to take advantage of streaming Netflix, Hulu, and the like. Instead of just having a stripped down Linux distro or Android fork with a web browser inside, they each had a different set of applications to hook up with popular streaming services. Each TV had their own Netflix app , Hulu App, Amazon Instant Video App, Youtube App, etc. Regardless of who developed these apps, be it Youtube or Sony, this sort of platform-specific application development absolutely kills innovation. How can we expect to get new and innovative ways to use our TV’s if every new entrant has to develop for so many different platforms, if they’re given access at all?
The web browser is magical - it gives every computing device access to every web app in the world. By not including a web browser, the simplest of features for an internet-connected device, these Smart TV’s are closing us off from the innovation we’ve become accustomed to.
As the hardware revolution brings computing to an increasing number of devices, we need to fight for an open web on those devices to keep software innovation moving forward.
 - Yes, there are differences in web browsers that have to be accounted for, and take valuable development time, but those are getting better with stronger standards, and they are nothing compared to developing for say, Windows and OSX and Linux.